
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.datasets import make_blobs
from sklearn.cluster import KMeans
from sklearn.preprocessing import StandardScaler
from sklearn.preprocessing import MinMaxScaler
from sklearn.decomposition import PCA
from scipy.cluster.hierarchy import dendrogram, linkage
from sklearn.cluster import AgglomerativeClustering
from sklearn.metrics import calinski_harabasz_score
from scipy.cluster.hierarchy import dendrogram
from scipy.cluster import hierarchy
import scipy.cluster.hierarchy as shc
from kneed import KneeLocator
from sklearn.metrics import silhouette_samples, silhouette_score
# Ignore harmless warnings
import warnings
warnings.filterwarnings("ignore")
%matplotlib inline
from statsmodels.stats.outliers_influence import variance_inflation_factor as ViF
from sklearn.neighbors import NearestCentroid


Produksi_Buah_2020_2022=pd.read_excel("Produksi Tanaman Buah-buahan 2020-2022 gabungan.xlsx")
Produksi_Buah_2020_2022=Produksi_Buah_2020_2022.set_index("Provinsi")


Produksi_Buah_2020_2022.head()


Produksi_Buah_2020_2022.describe()


Produksi_Buah_2020_2022.info()

<class 'pandas.core.frame.DataFrame'>
Index: 34 entries, ACEH to PAPUA
Data columns (total 26 columns):
 #   Column       Non-Null Count  Dtype  
---  ------       --------------  -----  
 0   Alpukat      34 non-null     int64  
 1   Belimbing    34 non-null     int64  
 2   Duku         34 non-null     int64  
 3   Durian       34 non-null     int64  
 4   Jambu Biji   34 non-null     int64  
 5   Jambu Air    34 non-null     int64  
 6   Jeruk Siam   34 non-null     int64  
 7   Jeruk Besar  34 non-null     int64  
 8   Mangga       34 non-null     int64  
 9   Manggis      33 non-null     float64
 10  Nangka       34 non-null     int64  
 11  Nanas        33 non-null     float64
 12  Pepaya       34 non-null     int64  
 13  Pisang       34 non-null     int64  
 14  Rambutan     34 non-null     int64  
 15  Salak        34 non-null     int64  
 16  Sawo         34 non-null     int64  
 17  Markisa      29 non-null     float64
 18  Sirsak       34 non-null     int64  
 19  Sukun        34 non-null     int64  
 20  Melon        34 non-null     int64  
 21  Semangka     33 non-null     float64
 22  Blewah       22 non-null     float64
 23  Apel         10 non-null     float64
 24  Anggur       26 non-null     float64
 25  Stroberi     20 non-null     float64
dtypes: float64(8), int64(18)
memory usage: 7.2+ KB


sns.set(rc={"figure.figsize":(10, 6)})
box2020_2022 = sns.boxplot(data=Produksi_Buah_2020_2022,orient="h")
plt.xlabel("Berat Hasil Produksi Buah (dari jutaan ton)")
plt.ylabel("Jenis Hasil Produksi Buah")
plt.show(box2020_2022)


ProduksiBuah20_22=Produksi_Buah_2020_2022.drop(columns=['Markisa','Blewah']) 
#kedua variabel tidak mempunyai produksi diantara 34 provinsi pada tahun 2021 sehingga variabel tidak digunakan.


ProduksiBuah20_22= ProduksiBuah20_22.fillna(0)


ProduksiBuah20_22


sns.relplot(data=ProduksiBuah20_22)

<seaborn.axisgrid.FacetGrid at 0x192996d8a90>


sns.heatmap(ProduksiBuah20_22.isnull(),yticklabels=False,cbar=False,cmap='viridis')

<AxesSubplot:ylabel='Provinsi'>


sns.heatmap(ProduksiBuah20_22.corr())

<AxesSubplot:>


sns.set(rc={"figure.figsize":(10, 6)})
box20_22 = sns.boxplot(data=ProduksiBuah20_22,orient="h")
plt.xlabel("Berat Hasil Produksi Buah (dari jutaan ton)")
plt.ylabel("Jenis Hasil Produksi Buah")
plt.show(box20_22)


ProduksiBuah20_22_copy = ProduksiBuah20_22.copy()


pb = ProduksiBuah20_22.to_numpy()


scaler = MinMaxScaler()
#scaler = StandardScaler()


ProduksiBuah20_22_scaled = scaler.fit_transform(ProduksiBuah20_22.to_numpy())
ProduksiBuah20_22_scaled = pd.DataFrame(ProduksiBuah20_22_scaled, columns=ProduksiBuah20_22.columns,index=ProduksiBuah20_22.index)


print("Scaled Dataset Using MinMaxScaler")
ProduksiBuah20_22_scaled

Scaled Dataset Using MinMaxScaler


sns.relplot(data=ProduksiBuah20_22_scaled)

<seaborn.axisgrid.FacetGrid at 0x192991f66d0>


sns.set(rc={"figure.figsize":(10, 6)})
box20_22 = sns.boxplot(data=ProduksiBuah20_22_scaled,orient="h")
plt.xlabel("Berat Hasil Produksi Buah (dari jutaan ton)")
plt.ylabel("Jenis Hasil Produksi Buah")
plt.show(box20_22)


def Calc_VIF(Data):
    #Calculating VIF
    VIF = pd.DataFrame()
    VIF['Variables'] = Data.columns
    VIF['VIF'] = [ViF(Data.values, i) for i in range(Data.shape[1])]
    return VIF


Calc_VIF(ProduksiBuah20_22_scaled)


ProduksiBuah20_22_scaled_cluster = ProduksiBuah20_22_scaled.copy()
psc = ProduksiBuah20_22_scaled.to_numpy()


pca = PCA(n_components=3)


principalComponents = pca.fit_transform(ProduksiBuah20_22_scaled_cluster)


print('Explained variation per principal component: {}'.format(pca.explained_variance_ratio_))

Explained variation per principal component: [0.65229876 0.086718   0.06117068]


principalDf = pd.DataFrame(data = principalComponents
             , columns = ['principal component 1', 'principal component 2','principal component 3'],index = ProduksiBuah20_22_scaled.index)
#,'principal component 4','principal component 5','principal component 6','principal component 7','principal component 8','principal component 9','principal component 10'


principalDf


plt.figure(figsize=(15,7))
plt.scatter(principalComponents[:,0],principalComponents[:,1],s = 50,cmap='rainbow')
plt.xlabel('First principal component')
plt.ylabel('Second Principal Component')

Text(0, 0.5, 'Second Principal Component')


print('Explained variation per principal component: {}'.format(pca.explained_variance_ratio_))

Explained variation per principal component: [0.65229876 0.086718   0.06117068]


evc = pca.explained_variance_ratio_.cumsum()
evc

array([0.65229876, 0.73901675, 0.80018744])


plt.ylabel('% Variance Explained')
plt.xlabel('# of Features')
plt.title('PCA Analysis')
plt.style.context('seaborn-whitegrid')
#plt.axhline(y = 0.8, color = 'r', linestyle = '-')

plt.plot(evc)

[<matplotlib.lines.Line2D at 0x1929a4d3550>]


print(pca.singular_values_)

[4.82349523 1.75870583 1.47710115]


loadings = pd.DataFrame(pca.components_.T, columns=['PC1', 'PC2','PC3'],index=['Alpukat (Ton)','Belimbing (Ton)','Duku (Ton)','Durian (Ton)','Jambu Biji (Ton)','Jambu Air (Ton)','Jeruk Siam (Ton)','Jeruk Besar (Ton)','Mangga (Ton)','Manggis (Ton)','Nangka (Ton)','Nenas (Ton)','Pepaya (Ton)','Pisang (Ton)','Rambutan (Ton)','Salak (Ton)','Sawo (Ton)','Sirsak (Ton)','Sukun (Ton)','Melon (Ton)','Semangka (Ton)','Apel (Ton)','Anggur (Ton)','Stroberi (Ton)'])
loadings
#,'PC4','PC5','PC6','PC7','PC8','PC9','PC10'


pc1_loadings = loadings.sort_values(by='PC1', ascending=False)[['PC1']]
pc1_loadings = pc1_loadings.reset_index()
pc1_loadings.columns = ['Attribute', 'CorrelationWithPC1']

plt.bar(x=pc1_loadings['Attribute'], height=pc1_loadings['CorrelationWithPC1'], color='#087E8B')
plt.title('PCA loading scores (first principal component)', size=20)
plt.xticks(rotation='vertical')
plt.show()


wcss = []
for i in range (1,11):
    kmeans_pca = KMeans(i)
    kmeans_pca.fit(principalComponents)
    wcss.append(kmeans_pca.inertia_)


wcss

[28.54098019819429,
 9.372392587904084,
 5.3572564298669425,
 3.2111782243268263,
 1.962483422058978,
 1.3226630825115648,
 0.9173946451958412,
 0.5985585401544481,
 0.45682420658711,
 0.3127598148093504]


y=wcss
x=range(1, len(y)+1)
kn = KneeLocator(x, y, curve='convex', direction='decreasing')
print(kn.knee)

3


plt.title('The Elbow Method')
plt.xlabel('Number of clusters')
plt.ylabel('Within-cluster Sum of Squares')
plt.plot(x, y, 'bx-')
plt.vlines(kn.knee, plt.ylim()[0], plt.ylim()[1], linestyles='dashed')

<matplotlib.collections.LineCollection at 0x1929a61bfd0>


kmeans3_i = KMeans(n_clusters=3, init='k-means++',n_init=10,max_iter=300,random_state=42)
kmeans3_i.fit(principalDf)
identified_clusters_Kmeans3_i = kmeans3_i.fit_predict(principalDf)
identified_clusters_Kmeans3_i

array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 2, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])


kmeans3_i.feature_names_in_

array(['principal component 1', 'principal component 2',
       'principal component 3'], dtype=object)


kmeans3_r = KMeans(n_clusters=3, init='random',n_init=10,max_iter=300,random_state=42)
kmeans3_r.fit(principalDf)
identified_clusters_Kmeans3_r = kmeans3_r.fit_predict(principalDf)
identified_clusters_Kmeans3_r

array([2, 0, 0, 2, 2, 2, 2, 0, 2, 2, 2, 1, 1, 2, 1, 2, 2, 0, 2, 2, 2, 2,
       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2])


kmeans4 = KMeans(n_clusters=4, init='k-means++',n_init=10,max_iter=300,random_state=42)
kmeans4.fit(principalDf)
identified_clusters_Kmeans4 = kmeans4.fit_predict(principalDf)
identified_clusters_Kmeans4

array([0, 3, 3, 0, 0, 0, 0, 3, 0, 0, 0, 1, 1, 0, 2, 0, 0, 3, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])


kmeans4_r = KMeans(n_clusters=4, init='random',n_init=10,max_iter=300,random_state=42)
kmeans4_r.fit(principalDf)
identified_clusters_Kmeans4_r = kmeans4_r.fit_predict(principalDf)
identified_clusters_Kmeans4_r

array([2, 1, 1, 2, 2, 2, 0, 1, 0, 0, 0, 1, 3, 2, 3, 2, 2, 1, 2, 2, 0, 0,
       0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0])


kmeans5 = KMeans(n_clusters=5, init='k-means++',n_init=10,max_iter=300,random_state=42)
kmeans5.fit(principalDf)
identified_clusters_Kmeans5 = kmeans5.fit_predict(principalDf)
identified_clusters_Kmeans5

array([2, 0, 0, 2, 2, 2, 2, 0, 2, 2, 2, 4, 1, 2, 3, 2, 2, 0, 2, 2, 2, 2,
       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2])


kmeans5_r = KMeans(n_clusters=5, init='random',n_init=10,max_iter=300,random_state=42)
kmeans5_r.fit(principalDf)
identified_clusters_Kmeans5_r = kmeans5_r.fit_predict(principalDf)
identified_clusters_Kmeans5_r

array([2, 1, 1, 2, 2, 2, 0, 1, 0, 0, 0, 4, 4, 2, 3, 2, 2, 1, 2, 2, 0, 0,
       0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0])


plt.scatter(principalComponents[:, 0], principalComponents[:, 1], c= identified_clusters_Kmeans3_i, s=100, cmap='rainbow')
plt.title("Optimal Number of Clusters")

centers = kmeans3_i.cluster_centers_
plt.scatter(centers[:, 0], centers[:, 1], c='black', s=200, alpha=0.5)
plt.xlabel('First principal component')
plt.ylabel('Second Principal Component')
plt.show()


plt.scatter(principalComponents[:, 0], principalComponents[:, 1], c= identified_clusters_Kmeans3_r, s=100, cmap='rainbow')
plt.title("Optimal Number of Clusters")

centers = kmeans3_r.cluster_centers_
plt.scatter(centers[:, 0], centers[:, 1], c='black', s=200, alpha=0.5)
plt.xlabel('First principal component')
plt.ylabel('Second Principal Component')
plt.show()


Kmeans_Tabel = pd.DataFrame(data =identified_clusters_Kmeans3_i, columns=['k=3 kmeans++'], index = ProduksiBuah20_22_scaled_cluster.index)


Kmeans_Tabel['k=4 kmeans++'] = identified_clusters_Kmeans4
Kmeans_Tabel['k=5 kmeans++'] = identified_clusters_Kmeans5
Kmeans_Tabel['k=3 random'] = identified_clusters_Kmeans3_r
Kmeans_Tabel['k=4 random'] = identified_clusters_Kmeans4_r
Kmeans_Tabel['k=5 random'] = identified_clusters_Kmeans5_r


Kmeans_Tabel


cluster3 = AgglomerativeClustering(n_clusters=3, affinity='euclidean', linkage='single')
cluster3 =cluster3.fit_predict(principalComponents)
cluster3

array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0, 1, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], dtype=int64)


cluster4 = AgglomerativeClustering(n_clusters=4, affinity='euclidean', linkage='single')
cluster4 =cluster4.fit_predict(principalComponents)
cluster4

array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 2, 0, 1, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], dtype=int64)


cluster5 = AgglomerativeClustering(n_clusters=5, affinity='euclidean', linkage='single')
cluster5 =cluster5.fit_predict(principalComponents)
cluster5

array([0, 0, 4, 0, 0, 0, 0, 0, 0, 0, 0, 3, 2, 0, 1, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], dtype=int64)


plt.figure(figsize=(15, 7))
plt.title("Dendrogram")
dend = dendrogram(hierarchy.linkage(principalComponents, method='single'))
plt.axhline(y=1, color='black', linestyle='--')
plt.show()


SingleLinkage_Tabel = pd.DataFrame(data =cluster3, columns=['k=3'], index = ProduksiBuah20_22_scaled_cluster.index)


SingleLinkage_Tabel['k=4'] = cluster4
SingleLinkage_Tabel['k=5'] = cluster5


SingleLinkage_Tabel


clf3 = NearestCentroid()
clf3.fit(principalComponents, cluster3 )
slc3=clf3.centroids_
print(slc3)

[[-0.17586741 -0.01699351 -0.02067896]
 [ 3.45358047  1.08084392 -0.17360785]
 [ 2.17417666 -0.5370515   0.83533448]]


clf4 = NearestCentroid()
clf4.fit(principalComponents, cluster4 )
slc4=clf4.centroids_
print(slc4)

[[-0.23303979  0.0154807  -0.00351477]
 [ 3.45358047  1.08084392 -0.17360785]
 [ 2.17417666 -0.5370515   0.83533448]
 [ 1.59647623 -1.02369398 -0.55276887]]


clf5 = NearestCentroid()
clf5.fit(principalComponents, cluster5 )
slc5=clf5.centroids_
print(slc5)

[[-0.25363976  0.02108769  0.02223899]
 [ 3.45358047  1.08084392 -0.17360785]
 [ 2.17417666 -0.5370515   0.83533448]
 [ 1.59647623 -1.02369398 -0.55276887]
 [ 0.38495947 -0.15272904 -0.77612743]]


slc3_ = np.array(slc3)
slc4_ = np.array(slc4)
slc5_ = np.array(slc5)


hkmeans3 = KMeans(n_clusters=3, init=slc3_,n_init=10,max_iter=300,random_state=42)
hkmeans3.fit(principalComponents)
identified_clusters_hKmeans3 = hkmeans3.fit_predict(principalComponents)
identified_clusters_hKmeans3

array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 0, 1, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])


hkmeans4 = KMeans(n_clusters=4, init=slc4_,n_init=10,max_iter=300,random_state=42)
hkmeans4.fit(principalComponents)
identified_clusters_hKmeans4 = hkmeans4.fit_predict(principalComponents)
identified_clusters_hKmeans4

array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 2, 0, 1, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])


hkmeans5 = KMeans(n_clusters=5, init=slc5_,n_init=10,max_iter=300,random_state=42)
hkmeans5.fit(principalComponents)
identified_clusters_hKmeans5 = hkmeans5.fit_predict(principalComponents)
identified_clusters_hKmeans5

array([0, 4, 4, 0, 0, 0, 0, 4, 0, 0, 0, 3, 2, 0, 1, 0, 0, 4, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])


Hierarchical_Kmeans_Tabel = pd.DataFrame(data =identified_clusters_hKmeans3, columns=['k=3 hk'], index = ProduksiBuah20_22_scaled_cluster.index)


Hierarchical_Kmeans_Tabel['k=4 hk'] = identified_clusters_hKmeans4
Hierarchical_Kmeans_Tabel['k=5 hk'] = identified_clusters_hKmeans5


Hierarchical_Kmeans_Tabel


ProduksiBuah20_22_copy = pd.concat([ProduksiBuah20_22_copy, Kmeans_Tabel,SingleLinkage_Tabel,Hierarchical_Kmeans_Tabel ], axis=1)
ProduksiBuah20_22_scaled_cluster = pd.concat([ProduksiBuah20_22_scaled_cluster, Kmeans_Tabel,SingleLinkage_Tabel,Hierarchical_Kmeans_Tabel  ], axis=1)


kmeans3 = KMeans(n_clusters=3, init='k-means++',n_init=10,max_iter=300,random_state=42)
kmeans3.fit(ProduksiBuah20_22_scaled)
kmeans4 = KMeans(n_clusters=4, init='k-means++',n_init=10,max_iter=300,random_state=42)
kmeans4.fit(ProduksiBuah20_22_scaled)
kmeans5 = KMeans(n_clusters=5, init='k-means++',n_init=10,max_iter=300,random_state=42)
kmeans5.fit(ProduksiBuah20_22_scaled)

KMeans(n_clusters=5, random_state=42)


resultsk = {}

for i in range(3,6):
    kmeansk = KMeans(n_clusters=i, init='k-means++',n_init=10,max_iter=300,random_state=42)
    labelsk = kmeansk.fit_predict(principalComponents)
    db_indexk = calinski_harabasz_score(psc, labelsk)
    resultsk.update({i: db_indexk})

resultsr = {}

for i in range(3,6):
    kmeansr = KMeans(n_clusters=i, init='random',n_init=10,max_iter=300,random_state=42)
    labelsr = kmeansr.fit_predict(principalComponents)
    db_indexr = calinski_harabasz_score(psc, labelsr)
    resultsr.update({i: db_indexr})

resultshk = {}

for i in range(1,2):
    kmeanshk3_ = KMeans(n_clusters=3, init=slc3_,n_init=10,max_iter=300,random_state=42)
    labelshk3_ = kmeanshk3_.fit_predict(principalComponents)
    db_indexhk3_ = calinski_harabasz_score(psc, labelshk3_)
    resultshk.update({3: db_indexhk3_})
    kmeanshk4_ = KMeans(n_clusters=4, init=slc4_,n_init=10,max_iter=300,random_state=42)
    labelshk4_ = kmeanshk4_.fit_predict(principalComponents)
    db_indexhk4_ = calinski_harabasz_score(psc, labelshk4_)
    resultshk.update({4: db_indexhk4_})
    kmeanshk5_ = KMeans(n_clusters=5, init=slc5_,n_init=10,max_iter=300,random_state=42)
    labelshk5_ = kmeanshk5_.fit_predict(principalComponents)
    db_indexhk5_ = calinski_harabasz_score(psc, labelshk5_)
    resultshk.update({5: db_indexhk5_})

resultssl = {}

for i in range(3,6):
    sl = AgglomerativeClustering(n_clusters=i, affinity='euclidean', linkage='single')
    labelssl = sl.fit_predict(principalComponents)
    db_indexsl = calinski_harabasz_score(psc, labelssl)
    resultssl.update({i: db_indexsl})


sns.set(rc={"figure.figsize":(10, 6)})
plt.plot(list(resultsk.keys()), list(resultsk.values()), label='K-Means++')
plt.plot(list(resultsr.keys()), list(resultsr.values()), label='Random')
plt.plot(list(resultshk.keys()), list(resultshk.values()), label='Hierarchical K-Means')
plt.plot(list(resultssl.keys()), list(resultssl.values()), label='Single Linkage')
plt.title('Calinski Harabasz Score 2020-2022')
plt.xlabel("Number of clusters")
plt.ylabel("Calinski-Harabasz Index")
plt.legend()
plt.show()


ch_index_kmeans3_i = calinski_harabasz_score(psc, Kmeans_Tabel['k=3 kmeans++'])
ch_index_kmeans4_i = calinski_harabasz_score(psc, Kmeans_Tabel['k=4 kmeans++'])
ch_index_kmeans5_i = calinski_harabasz_score(psc, Kmeans_Tabel['k=5 kmeans++'])
ch_index_kmeans3_r = calinski_harabasz_score(psc, Kmeans_Tabel['k=3 random'])
ch_index_kmeans4_r = calinski_harabasz_score(psc, Kmeans_Tabel['k=4 random'])
ch_index_kmeans5_r = calinski_harabasz_score(psc, Kmeans_Tabel['k=5 random'])
ch_index_hkmeans3 = calinski_harabasz_score(psc, Hierarchical_Kmeans_Tabel['k=3 hk'])
ch_index_hkmeans4 = calinski_harabasz_score(psc, Hierarchical_Kmeans_Tabel['k=4 hk'])
ch_index_hkmeans5 = calinski_harabasz_score(psc, Hierarchical_Kmeans_Tabel['k=5 hk'])


print('Calinski Harabasz Index K-means 3 kmeans++:',ch_index_kmeans3_i)
print('Calinski Harabasz Index K-means 4 kmeans++:',ch_index_kmeans4_i)
print('Calinski Harabasz Index K-means 5 kmeans++:',ch_index_kmeans5_i)
print('Calinski Harabasz Index K-means 3 random:',ch_index_kmeans3_r)
print('Calinski Harabasz Index K-means 3 random:',ch_index_kmeans4_r)
print('Calinski Harabasz Index K-means 3 random:',ch_index_kmeans5_r)
print('Calinski Harabasz Index HK-means 3:',ch_index_hkmeans3)
print('Calinski Harabasz Index HK-means 4:',ch_index_hkmeans4)
print('Calinski Harabasz Index HK-means 5:',ch_index_hkmeans5)

Calinski Harabasz Index K-means 3 kmeans++: 30.27567892886422
Calinski Harabasz Index K-means 4 kmeans++: 28.756485814113994
Calinski Harabasz Index K-means 5 kmeans++: 27.523183992221334
Calinski Harabasz Index K-means 3 random: 25.64022381449592
Calinski Harabasz Index K-means 3 random: 19.29857643823079
Calinski Harabasz Index K-means 3 random: 24.12419255437759
Calinski Harabasz Index HK-means 3: 30.27567892886422
Calinski Harabasz Index HK-means 4: 24.59244988312941
Calinski Harabasz Index HK-means 5: 27.523183992221334


ch_index_SingleLinkage3 = calinski_harabasz_score(psc, cluster3 )
ch_index_SingleLinkage4 = calinski_harabasz_score(psc, cluster4 )
ch_index_SingleLinkage5 = calinski_harabasz_score(psc, cluster5 )


print('Calinski Harabasz Index Single Linkage 3:',ch_index_SingleLinkage3)
print('Calinski Harabasz Index Single Linkage 4:',ch_index_SingleLinkage4)
print('Calinski Harabasz Index Single Linkage 5:',ch_index_SingleLinkage5)

Calinski Harabasz Index Single Linkage 3: 20.493703416896846
Calinski Harabasz Index Single Linkage 4: 24.59244988312941
Calinski Harabasz Index Single Linkage 5: 22.101845709584115


resultsk = {}

for i in range(3,6):
    kmeansk = KMeans(n_clusters=i, init='k-means++',n_init=10,max_iter=300,random_state=42)
    labelsk = kmeansk.fit_predict(principalComponents)
    db_indexk = silhouette_score(psc, labelsk)
    resultsk.update({i: db_indexk})

resultsr = {}

for i in range(3,6):
    kmeansr = KMeans(n_clusters=i, init='random',n_init=10,max_iter=300,random_state=42)
    labelsr = kmeansr.fit_predict(principalComponents)
    db_indexr = silhouette_score(psc, labelsr)
    resultsr.update({i: db_indexr})

resultshk = {}

for i in range(1,2):
    kmeanshk3_ = KMeans(n_clusters=3, init=slc3_,n_init=10,max_iter=300,random_state=42)
    labelshk3_ = kmeanshk3_.fit_predict(principalComponents)
    db_indexhk3_ = silhouette_score(psc, labelshk3_)
    resultshk.update({3: db_indexhk3_})
    kmeanshk4_ = KMeans(n_clusters=4, init=slc4_,n_init=10,max_iter=300,random_state=42)
    labelshk4_ = kmeanshk4_.fit_predict(principalComponents)
    db_indexhk4_ = silhouette_score(psc, labelshk4_)
    resultshk.update({4: db_indexhk4_})
    kmeanshk5_ = KMeans(n_clusters=5, init=slc5_,n_init=10,max_iter=300,random_state=42)
    labelshk5_ = kmeanshk5_.fit_predict(principalComponents)
    db_indexhk5_ = silhouette_score(psc, labelshk5_)
    resultshk.update({5: db_indexhk5_})

resultssl = {}

for i in range(3,6):
    sl = AgglomerativeClustering(n_clusters=i, affinity='euclidean', linkage='single')
    labelssl = sl.fit_predict(principalComponents)
    db_indexsl = silhouette_score(psc, labelssl)
    resultssl.update({i: db_indexsl})


sns.set(rc={"figure.figsize":(10, 6)})
plt.plot(list(resultsk.keys()), list(resultsk.values()), label='K-Means++')
plt.plot(list(resultsr.keys()), list(resultsr.values()), label='Random')
plt.plot(list(resultshk.keys()), list(resultshk.values()), label='Hierarchical K-Means')
plt.plot(list(resultssl.keys()), list(resultssl.values()), label='Single Linkage')
plt.title('Silhouette Coefficient Score 2020-2022')
plt.xlabel("Number of clusters")
plt.ylabel("Silhouette Coefficient Index")
plt.legend()
plt.show()


score3_i = silhouette_score(psc, Kmeans_Tabel['k=3 kmeans++'], metric='euclidean')
score4_i = silhouette_score(psc, Kmeans_Tabel['k=4 kmeans++'], metric='euclidean')
score5_i = silhouette_score(psc, Kmeans_Tabel['k=5 kmeans++'], metric='euclidean')
score3_r = silhouette_score(psc, Kmeans_Tabel['k=3 random'], metric='euclidean')
score4_r = silhouette_score(psc, Kmeans_Tabel['k=4 random'], metric='euclidean')
score5_r = silhouette_score(psc, Kmeans_Tabel['k=5 random'], metric='euclidean')
hscore3 = silhouette_score(psc, identified_clusters_hKmeans3, metric='euclidean')
hscore4 = silhouette_score(psc, identified_clusters_hKmeans4, metric='euclidean')
hscore5 = silhouette_score(psc, identified_clusters_hKmeans5, metric='euclidean')


print('Silhouette coefficient K-means 3 k-means++:',score3_i)
print('Silhouette coefficient K-means 4 k-means++:',score4_i)
print('Silhouette coefficient K-means 5 k-means++:',score5_i)
print('Silhouette coefficient K-means 3 random:',score3_r)
print('Silhouette coefficient K-means 4 random:',score4_r)
print('Silhouette coefficient K-means 5 random:',score5_r)
print('Silhouette coefficient HK-means 3:',hscore3)
print('Silhouette coefficient HK-means 4:',hscore4)
print('Silhouette coefficient HK-means 5:',hscore5)

Silhouette coefficient K-means 3 k-means++: 0.6745383942607401
Silhouette coefficient K-means 4 k-means++: 0.5024000913928874
Silhouette coefficient K-means 5 k-means++: 0.4941079464595051
Silhouette coefficient K-means 3 random: 0.4991462727828761
Silhouette coefficient K-means 4 random: 0.2957044656981662
Silhouette coefficient K-means 5 random: 0.3073022491918449
Silhouette coefficient HK-means 3: 0.6745383942607401
Silhouette coefficient HK-means 4: 0.6408219489733553
Silhouette coefficient HK-means 5: 0.4941079464595051


scores3 = silhouette_score(psc, cluster3, metric='euclidean')
scores4 = silhouette_score(psc, cluster4, metric='euclidean')
scores5 = silhouette_score(psc, cluster5, metric='euclidean')


print('Silhouette coefficient K-means 3:',scores3)
print('Silhouette coefficient K-means 4:',scores4)
print('Silhouette coefficient K-means 5:',scores5)

Silhouette coefficient K-means 3: 0.6560560906881828
Silhouette coefficient K-means 4: 0.6408219489733553
Silhouette coefficient K-means 5: 0.47206141916418703


ch_index_kmeans3_i = calinski_harabasz_score(pb, Kmeans_Tabel['k=3 kmeans++'])
ch_index_kmeans4_i = calinski_harabasz_score(pb, Kmeans_Tabel['k=4 kmeans++'])
ch_index_kmeans5_i = calinski_harabasz_score(pb, Kmeans_Tabel['k=5 kmeans++'])
ch_index_kmeans3_r = calinski_harabasz_score(pb, Kmeans_Tabel['k=3 random'])
ch_index_kmeans4_r = calinski_harabasz_score(pb, Kmeans_Tabel['k=4 random'])
ch_index_kmeans5_r = calinski_harabasz_score(pb, Kmeans_Tabel['k=5 random'])
ch_index_hkmeans3 = calinski_harabasz_score(pb, Hierarchical_Kmeans_Tabel['k=3 hk'])
ch_index_hkmeans4 = calinski_harabasz_score(pb, Hierarchical_Kmeans_Tabel['k=4 hk'])
ch_index_hkmeans5 = calinski_harabasz_score(pb, Hierarchical_Kmeans_Tabel['k=5 hk'])


print('Calinski Harabasz Index K-means 3 kmeans++:',ch_index_kmeans3_i)
print('Calinski Harabasz Index K-means 4 kmeans++:',ch_index_kmeans4_i)
print('Calinski Harabasz Index K-means 5 kmeans++:',ch_index_kmeans5_i)
print('Calinski Harabasz Index K-means 3 random:',ch_index_kmeans3_r)
print('Calinski Harabasz Index K-means 3 random:',ch_index_kmeans4_r)
print('Calinski Harabasz Index K-means 3 random:',ch_index_kmeans5_r)
print('Calinski Harabasz Index HK-means 3:',ch_index_hkmeans3)
print('Calinski Harabasz Index HK-means 4:',ch_index_hkmeans4)
print('Calinski Harabasz Index HK-means 5:',ch_index_hkmeans5)

Calinski Harabasz Index K-means 3 kmeans++: 57.982645118375565
Calinski Harabasz Index K-means 4 kmeans++: 49.69512635670209
Calinski Harabasz Index K-means 5 kmeans++: 41.93869884583579
Calinski Harabasz Index K-means 3 random: 29.186360201480685
Calinski Harabasz Index K-means 3 random: 16.58936536079578
Calinski Harabasz Index K-means 3 random: 39.53415387813175
Calinski Harabasz Index HK-means 3: 57.98264511837555
Calinski Harabasz Index HK-means 4: 42.40873814090654
Calinski Harabasz Index HK-means 5: 41.93869884583579


ch_index_SingleLinkage3 = calinski_harabasz_score(psc, cluster3 )
ch_index_SingleLinkage4 = calinski_harabasz_score(psc, cluster4 )
ch_index_SingleLinkage5 = calinski_harabasz_score(psc, cluster5 )


print('Calinski Harabasz Index Single Linkage 3:',ch_index_SingleLinkage3)
print('Calinski Harabasz Index Single Linkage 4:',ch_index_SingleLinkage4)
print('Calinski Harabasz Index Single Linkage 5:',ch_index_SingleLinkage5)

Calinski Harabasz Index Single Linkage 3: 20.493703416896846
Calinski Harabasz Index Single Linkage 4: 24.59244988312941
Calinski Harabasz Index Single Linkage 5: 22.101845709584115


score3_i = silhouette_score(pb, Kmeans_Tabel['k=3 kmeans++'], metric='euclidean')
score4_i = silhouette_score(pb, Kmeans_Tabel['k=4 kmeans++'], metric='euclidean')
score5_i = silhouette_score(pb, Kmeans_Tabel['k=5 kmeans++'], metric='euclidean')
score3_r = silhouette_score(pb, Kmeans_Tabel['k=3 random'], metric='euclidean')
score4_r = silhouette_score(pb, Kmeans_Tabel['k=4 random'], metric='euclidean')
score5_r = silhouette_score(pb, Kmeans_Tabel['k=5 random'], metric='euclidean')
hscore3 = silhouette_score(pb, identified_clusters_hKmeans3, metric='euclidean')
hscore4 = silhouette_score(pb, identified_clusters_hKmeans4, metric='euclidean')
hscore5 = silhouette_score(pb, identified_clusters_hKmeans5, metric='euclidean')


print('Silhouette coefficient K-means 3 k-means++:',score3_i)
print('Silhouette coefficient K-means 4 k-means++:',score4_i)
print('Silhouette coefficient K-means 5 k-means++:',score5_i)
print('Silhouette coefficient K-means 3 random:',score3_r)
print('Silhouette coefficient K-means 4 random:',score4_r)
print('Silhouette coefficient K-means 5 random:',score5_r)
print('Silhouette coefficient HK-means 3:',hscore3)
print('Silhouette coefficient HK-means 4:',hscore4)
print('Silhouette coefficient HK-means 5:',hscore5)

Silhouette coefficient K-means 3 k-means++: 0.726621062844469
Silhouette coefficient K-means 4 k-means++: 0.5201142248204489
Silhouette coefficient K-means 5 k-means++: 0.4986093315415186
Silhouette coefficient K-means 3 random: 0.5105054377850797
Silhouette coefficient K-means 4 random: 0.23293788852406308
Silhouette coefficient K-means 5 random: 0.25895015082520845
Silhouette coefficient HK-means 3: 0.726621062844469
Silhouette coefficient HK-means 4: 0.6751365567576909
Silhouette coefficient HK-means 5: 0.4986093315415186


scores3 = silhouette_score(pb, cluster3, metric='euclidean')
scores4 = silhouette_score(pb, cluster4, metric='euclidean')
scores5 = silhouette_score(pb, cluster5, metric='euclidean')


print('Silhouette coefficient K-means 3:',scores3)
print('Silhouette coefficient K-means 4:',scores4)
print('Silhouette coefficient K-means 5:',scores5)

Silhouette coefficient K-means 3: 0.6340872807029864
Silhouette coefficient K-means 4: 0.6751365567576909
Silhouette coefficient K-means 5: 0.12120922846162763


class Radar(object):
    def __init__(self, figure, title, labels, rect=None):
        if rect is None:
            rect = [0.05, 0.05, 0.9, 0.9]

        self.n = len(title)
        self.angles = np.arange(0, 360, 360.0/self.n)
        
        self.axes = [figure.add_axes(rect, projection='polar', label='axes%d' % i) for i in range(self.n)]
        self.ax = self.axes[0]
        self.ax.set_thetagrids(self.angles, labels=title, fontsize=14, backgroundcolor="white",zorder=999) # Feature names
        self.ax.set_yticklabels([])
        
        for ax in self.axes[1:]:
            ax.xaxis.set_visible(False)
            ax.set_yticklabels([])
            ax.set_zorder(-99)

        for ax, angle, label in zip(self.axes, self.angles, labels):
            ax.spines['polar'].set_color('black')
            ax.spines['polar'].set_zorder(-99)
                     
    def plot(self, values, *args, **kw):
        angle = np.deg2rad(np.r_[self.angles, self.angles[0]])
        values = np.r_[values, values[0]]
        self.ax.plot(angle, values, *args, **kw)
        kw['label'] = '_noLabel'
        self.ax.fill(angle, values,*args,**kw)


# Let's define max. 6 different cluster colors - if you like you can add here more.
cluster_colors = ['#b4d2b1', '#568f8b', '#1d4a60', '#cd7e59', '#ddb247', '#d15252']


X_mean3_i = pd.concat([pd.DataFrame(ProduksiBuah20_22_copy.mean(), columns=['mean']), 
                   ProduksiBuah20_22_copy.groupby('k=3 kmeans++').mean().T], axis=1)
X_dev_rel3_i = X_mean3_i.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_dev_rel3_i.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_mean3_i.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)

X_std_mean3_i = pd.concat([pd.DataFrame(ProduksiBuah20_22_scaled_cluster.mean(), columns=['mean']), 
                   ProduksiBuah20_22_scaled_cluster.groupby('k=3 kmeans++').mean().T], axis=1)

X_std_dev_rel3_i = X_std_mean3_i.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_std_dev_rel3_i.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_std_mean3_i.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)


fig = plt.figure(figsize=(8, 8))
no_features = len(kmeans3.feature_names_in_)
radar = Radar(fig, kmeans3.feature_names_in_, np.unique(kmeans3.labels_))

for k in range(0,kmeans3.n_clusters):
    cluster_data = X_std_mean3_i[k].values.tolist()
    radar.plot(cluster_data, '-', lw=2, color=cluster_colors[k], alpha=0.7, label='cluster {}'.format(k))

radar.ax.legend()
radar.ax.set_title("Cluster 3 kmeans++: Feature means per cluster", size=22, pad=60)
plt.show()


X_mean4_i = pd.concat([pd.DataFrame(ProduksiBuah20_22_copy.mean(), columns=['mean']), 
                   ProduksiBuah20_22_copy.groupby('k=4 kmeans++').mean().T], axis=1)
X_dev_rel4_i = X_mean4_i.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_dev_rel4_i.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_mean4_i.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)

X_std_mean4_i = pd.concat([pd.DataFrame(ProduksiBuah20_22_scaled_cluster.mean(), columns=['mean']), 
                   ProduksiBuah20_22_scaled_cluster.groupby('k=4 kmeans++').mean().T], axis=1)

X_std_dev_rel4_i = X_std_mean4_i.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_std_dev_rel4_i.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_std_mean4_i.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)


fig = plt.figure(figsize=(8, 8))
no_features = len(kmeans4.feature_names_in_)
radar = Radar(fig, kmeans4.feature_names_in_, np.unique(kmeans3.labels_))

for k in range(0,kmeans4.n_clusters):
    cluster_data = X_std_mean4_i[k].values.tolist()
    radar.plot(cluster_data, '-', lw=2, color=cluster_colors[k], alpha=0.7, label='cluster {}'.format(k))

radar.ax.legend()
radar.ax.set_title("Cluster 4 kmeans++: Feature means per cluster", size=22, pad=60)
plt.show()


X_mean5_i = pd.concat([pd.DataFrame(ProduksiBuah20_22_copy.mean(), columns=['mean']), 
                   ProduksiBuah20_22_copy.groupby('k=5 kmeans++').mean().T], axis=1)
X_dev_rel5_i = X_mean5_i.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_dev_rel5_i.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_mean5_i.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)

X_std_mean5_i = pd.concat([pd.DataFrame(ProduksiBuah20_22_scaled_cluster.mean(), columns=['mean']), 
                   ProduksiBuah20_22_scaled_cluster.groupby('k=5 kmeans++').mean().T], axis=1)

X_std_dev_rel5_i = X_std_mean5_i.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_std_dev_rel5_i.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_std_mean5_i.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)


fig = plt.figure(figsize=(8, 8))
no_features = len(kmeans5.feature_names_in_)
radar = Radar(fig, kmeans5.feature_names_in_, np.unique(kmeans5.labels_))

for k in range(0,kmeans5.n_clusters):
    cluster_data = X_std_mean5_i[k].values.tolist()
    radar.plot(cluster_data, '-', lw=2, color=cluster_colors[k], alpha=0.7, label='cluster {}'.format(k))

radar.ax.legend()
radar.ax.set_title("Cluster 5 kmeans++: Feature means per cluster", size=22, pad=60)
plt.show()


X_mean3_r = pd.concat([pd.DataFrame(ProduksiBuah20_22_copy.mean(), columns=['mean']), 
                   ProduksiBuah20_22_copy.groupby('k=3 random').mean().T], axis=1)
X_dev_rel3_r = X_mean3_r.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_dev_rel3_r.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_mean3_r.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)

X_std_mean3_r = pd.concat([pd.DataFrame(ProduksiBuah20_22_scaled_cluster.mean(), columns=['mean']), 
                   ProduksiBuah20_22_scaled_cluster.groupby('k=3 random').mean().T], axis=1)

X_std_dev_rel3_r = X_std_mean3_r.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_std_dev_rel3_r.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_std_mean3_r.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)


fig = plt.figure(figsize=(8, 8))
no_features = len(kmeans3.feature_names_in_)
radar = Radar(fig, kmeans3.feature_names_in_, np.unique(kmeans3.labels_))

for k in range(0,kmeans3.n_clusters):
    cluster_data = X_std_mean3_r[k].values.tolist()
    radar.plot(cluster_data, '-', lw=2, color=cluster_colors[k], alpha=0.7, label='cluster {}'.format(k))

radar.ax.legend()
radar.ax.set_title("Cluster 3 random: Feature means per cluster", size=22, pad=60)
plt.show()


X_mean4_r = pd.concat([pd.DataFrame(ProduksiBuah20_22_copy.mean(), columns=['mean']), 
                   ProduksiBuah20_22_copy.groupby('k=4 random').mean().T], axis=1)
X_dev_rel4_r = X_mean4_r.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_dev_rel4_r.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_mean4_r.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)

X_std_mean4_r = pd.concat([pd.DataFrame(ProduksiBuah20_22_scaled_cluster.mean(), columns=['mean']), 
                   ProduksiBuah20_22_scaled_cluster.groupby('k=4 random').mean().T], axis=1)

X_std_dev_rel4_r = X_std_mean4_r.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_std_dev_rel4_r.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_std_mean4_r.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)


fig = plt.figure(figsize=(8, 8))
no_features = len(kmeans4.feature_names_in_)
radar = Radar(fig, kmeans4.feature_names_in_, np.unique(kmeans4.labels_))

for k in range(0,kmeans4.n_clusters):
    cluster_data = X_std_mean4_r[k].values.tolist()
    radar.plot(cluster_data, '-', lw=2, color=cluster_colors[k], alpha=0.7, label='cluster {}'.format(k))

radar.ax.legend()
radar.ax.set_title("Cluster 4 random: Feature means per cluster", size=22, pad=60)
plt.show()


X_mean5_r = pd.concat([pd.DataFrame(ProduksiBuah20_22_copy.mean(), columns=['mean']), 
                   ProduksiBuah20_22_copy.groupby('k=5 random').mean().T], axis=1)
X_dev_rel5_r = X_mean5_r.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_dev_rel5_r.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_mean5_r.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)

X_std_mean5_r = pd.concat([pd.DataFrame(ProduksiBuah20_22_scaled_cluster.mean(), columns=['mean']), 
                   ProduksiBuah20_22_scaled_cluster.groupby('k=5 random').mean().T], axis=1)

X_std_dev_rel5_r = X_std_mean5_r.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_std_dev_rel5_r.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_std_mean5_r.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)


fig = plt.figure(figsize=(8, 8))
no_features = len(kmeans5.feature_names_in_)
radar = Radar(fig, kmeans5.feature_names_in_, np.unique(kmeans5.labels_))

for k in range(0,kmeans5.n_clusters):
    cluster_data = X_std_mean5_r[k].values.tolist()
    radar.plot(cluster_data, '-', lw=2, color=cluster_colors[k], alpha=0.7, label='cluster {}'.format(k))

radar.ax.legend()
radar.ax.set_title("Cluster 5 random: Feature means per cluster", size=22, pad=60)
plt.show()


X_mean3_hk = pd.concat([pd.DataFrame(ProduksiBuah20_22_copy.mean(), columns=['mean']), 
                   ProduksiBuah20_22_copy.groupby('k=3 hk').mean().T], axis=1)
X_dev_rel3_hk = X_mean3_hk.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_dev_rel3_hk.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_mean3_hk.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)

X_std_mean3_hk = pd.concat([pd.DataFrame(ProduksiBuah20_22_scaled_cluster.mean(), columns=['mean']), 
                   ProduksiBuah20_22_scaled_cluster.groupby('k=3 hk').mean().T], axis=1)

X_std_dev_rel3_hk = X_std_mean3_hk.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_std_dev_rel3_hk.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_std_mean3_hk.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)


fig = plt.figure(figsize=(8, 8))
no_features = len(kmeans3.feature_names_in_)
radar = Radar(fig, kmeans3.feature_names_in_, np.unique(kmeans3.labels_))

for k in range(0,kmeans3.n_clusters):
    cluster_data = X_std_mean3_hk[k].values.tolist()
    radar.plot(cluster_data, '-', lw=2, color=cluster_colors[k], alpha=0.7, label='cluster {}'.format(k))

radar.ax.legend()
radar.ax.set_title("Cluster 3 hierarchical kmeans: Feature means per cluster", size=22, pad=60)
plt.show()


X_mean4_hk = pd.concat([pd.DataFrame(ProduksiBuah20_22_copy.mean(), columns=['mean']), 
                   ProduksiBuah20_22_copy.groupby('k=4 hk').mean().T], axis=1)
X_dev_rel4_hk = X_mean4_hk.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_dev_rel4_hk.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_mean4_hk.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)

X_std_mean4_hk = pd.concat([pd.DataFrame(ProduksiBuah20_22_scaled_cluster.mean(), columns=['mean']), 
                   ProduksiBuah20_22_scaled_cluster.groupby('k=4 hk').mean().T], axis=1)

X_std_dev_rel4_hk = X_std_mean4_hk.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_std_dev_rel4_hk.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_std_mean4_hk.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)


fig = plt.figure(figsize=(8, 8))
no_features = len(kmeans4.feature_names_in_)
radar = Radar(fig, kmeans4.feature_names_in_, np.unique(kmeans4.labels_))

for k in range(0,kmeans4.n_clusters):
    cluster_data = X_std_mean4_hk[k].values.tolist()
    radar.plot(cluster_data, '-', lw=2, color=cluster_colors[k], alpha=0.7, label='cluster {}'.format(k))

radar.ax.legend()
radar.ax.set_title("Cluster 4 hierarchical kmeans: Feature means per cluster", size=22, pad=60)
plt.show()


X_mean5_hk = pd.concat([pd.DataFrame(ProduksiBuah20_22_copy.mean(), columns=['mean']), 
                   ProduksiBuah20_22_copy.groupby('k=5 hk').mean().T], axis=1)
X_dev_rel5_hk = X_mean5_hk.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_dev_rel5_hk.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_mean5_hk.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)

X_std_mean5_hk = pd.concat([pd.DataFrame(ProduksiBuah20_22_scaled_cluster.mean(), columns=['mean']), 
                   ProduksiBuah20_22_scaled_cluster.groupby('k=5 hk').mean().T], axis=1)

X_std_dev_rel5_hk = X_std_mean5_hk.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_std_dev_rel5_hk.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_std_mean5_hk.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)


fig = plt.figure(figsize=(8, 8))
no_features = len(kmeans5.feature_names_in_)
radar = Radar(fig, kmeans5.feature_names_in_, np.unique(kmeans5.labels_))

for k in range(0,kmeans5.n_clusters):
    cluster_data = X_std_mean5_hk[k].values.tolist()
    radar.plot(cluster_data, '-', lw=2, color=cluster_colors[k], alpha=0.7, label='cluster {}'.format(k))

radar.ax.legend()
radar.ax.set_title("Cluster 5 hierarchical kmeans: Feature means per cluster", size=22, pad=60)
plt.show()


X_mean3_sl = pd.concat([pd.DataFrame(ProduksiBuah20_22_copy.mean(), columns=['mean']), 
                   ProduksiBuah20_22_copy.groupby('k=3').mean().T], axis=1)
X_dev_rel3_sl = X_mean3_sl.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_dev_rel3_sl.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_mean3_sl.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)

X_std_mean3_sl = pd.concat([pd.DataFrame(ProduksiBuah20_22_scaled_cluster.mean(), columns=['mean']), 
                   ProduksiBuah20_22_scaled_cluster.groupby('k=3').mean().T], axis=1)

X_std_dev_rel3_sl = X_std_mean3_sl.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_std_dev_rel3_sl.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_std_mean3_sl.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)


fig = plt.figure(figsize=(8, 8))
no_features = len(kmeans3.feature_names_in_)
radar = Radar(fig, kmeans3.feature_names_in_, np.unique(kmeans3.labels_))

for k in range(0,kmeans3.n_clusters):
    cluster_data = X_std_mean3_sl[k].values.tolist()
    radar.plot(cluster_data, '-', lw=2, color=cluster_colors[k], alpha=0.7, label='cluster {}'.format(k))

radar.ax.legend()
radar.ax.set_title("Cluster 3 Single Linkage: Feature means per cluster", size=22, pad=60)
plt.show()


X_mean4_sl = pd.concat([pd.DataFrame(ProduksiBuah20_22_copy.mean(), columns=['mean']), 
                   ProduksiBuah20_22_copy.groupby('k=4').mean().T], axis=1)
X_dev_rel4_sl = X_mean4_sl.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_dev_rel4_sl.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_mean4_sl.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)

X_std_mean4_sl = pd.concat([pd.DataFrame(ProduksiBuah20_22_scaled_cluster.mean(), columns=['mean']), 
                   ProduksiBuah20_22_scaled_cluster.groupby('k=4').mean().T], axis=1)

X_std_dev_rel4_sl = X_std_mean4_sl.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_std_dev_rel4_sl.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_std_mean4_sl.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)


fig = plt.figure(figsize=(8, 8))
no_features = len(kmeans4.feature_names_in_)
radar = Radar(fig, kmeans4.feature_names_in_, np.unique(kmeans4.labels_))

for k in range(0,kmeans4.n_clusters):
    cluster_data = X_std_mean4_sl[k].values.tolist()
    radar.plot(cluster_data, '-', lw=2, color=cluster_colors[k], alpha=0.7, label='cluster {}'.format(k))

radar.ax.legend()
radar.ax.set_title("Cluster 4 Single Linkage: Feature means per cluster", size=22, pad=60)
plt.show()


X_mean5_sl = pd.concat([pd.DataFrame(ProduksiBuah20_22_copy.mean(), columns=['mean']), 
                   ProduksiBuah20_22_copy.groupby('k=5').mean().T], axis=1)
X_dev_rel5_sl = X_mean5_sl.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_dev_rel5_sl.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_mean5_sl.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)

X_std_mean5_sl = pd.concat([pd.DataFrame(ProduksiBuah20_22_scaled_cluster.mean(), columns=['mean']), 
                   ProduksiBuah20_22_scaled_cluster.groupby('k=5').mean().T], axis=1)

X_std_dev_rel5_sl = X_std_mean5_sl.apply(lambda x: round((x-x['mean'])/x['mean'],2)*100, axis = 1)
X_std_dev_rel5_sl.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)
X_std_mean5_sl.drop(columns=['mean'],index=['k=3 kmeans++','k=4 kmeans++','k=5 kmeans++','k=3 random','k=4 random','k=5 random','k=3','k=4','k=5','k=3 hk','k=4 hk','k=5 hk'], inplace=True)


fig = plt.figure(figsize=(8, 8))
no_features = len(kmeans5.feature_names_in_)
radar = Radar(fig, kmeans5.feature_names_in_, np.unique(kmeans5.labels_))

for k in range(0,kmeans5.n_clusters):
    cluster_data = X_std_mean5_sl[k].values.tolist()
    radar.plot(cluster_data, '-', lw=2, color=cluster_colors[k], alpha=0.7, label='cluster {}'.format(k))

radar.ax.legend()
radar.ax.set_title("Cluster 5 Single Linkage: Feature means per cluster", size=22, pad=60)
plt.show()

	Alpukat	Belimbing	Duku	Durian	Jambu Biji	Jambu Air	Jeruk Siam	Jeruk Besar	Mangga	Manggis	...	Sawo	Markisa	Sirsak	Sukun	Melon	Semangka	Blewah	Apel	Anggur	Stroberi
Provinsi
ACEH	150887	4309	22393	130595	23115	14949	46396	30986	79170	8516.0	...	13592	1084.0	3505	5122	1084	21116.0	NaN	301.0	NaN	143.0
SUMATERA UTARA	117901	19347	58761	313678	40725	48230	1183180	5245	127630	71314.0	...	62409	837.0	9001	6314	5126	112944.0	1.0	NaN	6.0	254.0
SUMATERA BARAT	266593	2950	20591	521775	28045	29095	381107	1684	49960	221084.0	...	36591	36500.0	27086	9692	5611	51088.0	20.0	47.0	4.0	438.0
RIAU	4579	6706	21868	75321	33291	28619	254484	4780	53924	30975.0	...	15694	6.0	4623	9492	3683	64243.0	58.0	NaN	59.0	NaN
JAMBI	33191	3579	50626	64455	11639	11071	168150	868	18281	30624.0	...	11771	84.0	6031	8945	2709	16897.0	52.0	NaN	12.0	21.0

	Alpukat	Belimbing	Duku	Durian	Jambu Biji	Jambu Air	Jeruk Siam	Jeruk Besar	Mangga	Manggis	...	Sawo	Markisa	Sirsak	Sukun	Melon	Semangka	Blewah	Apel	Anggur	Stroberi
count	34.000000	34.000000	34.000000	34.000000	34.000000	34.000000	3.400000e+01	34.000000	3.400000e+01	33.000000	...	34.000000	29.000000	34.000000	34.000000	34.000000	33.000000	22.000000	1.000000e+01	26.000000	20.00000
mean	63061.588235	11194.205882	21918.029412	119659.000000	37983.705882	18438.058824	2.219542e+05	11039.441176	2.659684e+05	29394.424242	...	15407.500000	1838.620690	12622.235294	15528.176471	11353.441176	40678.151515	1502.636364	1.549668e+05	1445.769231	2355.30000
std	117636.774955	35054.048100	27213.153865	194454.880051	77259.897458	31456.862337	4.903489e+05	26544.667640	7.499428e+05	52298.182488	...	22026.930314	6887.045896	23906.914817	25926.683755	34873.462228	71321.659603	4394.608109	4.895717e+05	6492.170449	8387.45517
min	120.000000	47.000000	27.000000	650.000000	66.000000	16.000000	6.490000e+02	45.000000	5.918000e+03	9.000000	...	1.000000	1.000000	102.000000	19.000000	10.000000	394.000000	1.000000	3.000000e+00	1.000000	1.00000
25%	2824.250000	1142.750000	4376.750000	20012.500000	3964.500000	3117.250000	1.445975e+04	610.500000	1.842575e+04	2399.000000	...	703.750000	8.000000	1256.250000	2347.000000	968.250000	4647.000000	10.250000	1.625000e+01	7.000000	23.75000
50%	10819.000000	2211.500000	14785.000000	55420.500000	11281.500000	5934.500000	5.845900e+04	1388.500000	4.531900e+04	6371.000000	...	5211.000000	70.000000	3878.000000	5535.000000	1903.500000	16897.000000	35.000000	5.750000e+01	27.500000	224.00000
75%	56122.500000	4250.500000	22849.750000	113428.250000	31145.500000	18469.500000	1.712295e+05	4477.500000	9.367225e+04	28751.000000	...	15168.500000	329.000000	8655.500000	9642.000000	4765.250000	47229.000000	168.750000	2.527500e+02	111.750000	452.50000
max	517574.000000	200209.000000	138448.000000	971027.000000	308606.000000	154531.000000	2.611658e+06	106417.000000	4.079350e+06	221084.000000	...	78460.000000	36500.000000	98601.000000	116228.000000	188639.000000	368149.000000	19709.000000	1.548312e+06	33207.000000	37826.00000

	Alpukat	Belimbing	Duku	Durian	Jambu Biji	Jambu Air	Jeruk Siam	Jeruk Besar	Mangga	Manggis	...	Rambutan	Salak	Sawo	Sirsak	Sukun	Melon	Semangka	Apel	Anggur	Stroberi
Provinsi
ACEH	150887	4309	22393	130595	23115	14949	46396	30986	79170	8516.0	...	100626	3034	13592	3505	5122	1084	21116.0	301.0	0.0	143.0
SUMATERA UTARA	117901	19347	58761	313678	40725	48230	1183180	5245	127630	71314.0	...	79792	859993	62409	9001	6314	5126	112944.0	0.0	6.0	254.0
SUMATERA BARAT	266593	2950	20591	521775	28045	29095	381107	1684	49960	221084.0	...	87406	6150	36591	27086	9692	5611	51088.0	47.0	4.0	438.0
RIAU	4579	6706	21868	75321	33291	28619	254484	4780	53924	30975.0	...	89349	12387	15694	4623	9492	3683	64243.0	0.0	59.0	0.0
JAMBI	33191	3579	50626	64455	11639	11071	168150	868	18281	30624.0	...	20499	17355	11771	6031	8945	2709	16897.0	0.0	12.0	21.0
SUMATERA SELATAN	105419	4068	138448	117258	18863	30010	172256	1184	52521	5916.0	...	40836	6044	21356	4299	8519	5271	45418.0	3.0	12.0	194.0
BENGKULU	22152	1089	2473	55868	5075	3337	69186	1370	22511	15944.0	...	18884	10062	5576	1413	1742	1368	2819.0	48.0	37.0	159.0
LAMPUNG	63732	10199	29695	62604	47611	31368	180335	3489	62078	28036.0	...	59123	26553	44535	8682	20771	2120	62769.0	0.0	90.0	1.0
KEP. BANGKA BELITUNG	4090	393	2103	12793	1745	4813	5472	279	5918	2399.0	...	5960	2998	1795	440	1558	1023	10017.0	0.0	0.0	0.0
KEP. RIAU	480	337	893	11568	1496	1037	698	69	6122	564.0	...	5060	20172	2717	1094	3385	277	5164.0	0.0	0.0	0.0
DKI JAKARTA	3423	7676	34	650	4554	6361	649	45	20108	0.0	...	11060	143	686	226	2506	10	0.0	0.0	14.0	0.0
JAWA BARAT	336037	25248	23002	240136	228644	78419	152813	23314	1337834	158602.0	...	436610	80852	78460	42522	75440	2505	29679.0	0.0	168.0	37826.0
JAWA TENGAH	279669	57908	45767	479012	308606	154531	84477	55588	1445394	28751.0	...	358116	1426748	63432	63961	116228	81932	203744.0	67.0	389.0	2898.0
DI YOGYAKARTA	24737	2274	8730	24878	15835	11786	7639	534	90355	8044.0	...	41298	166790	8678	8576	38525	43781	20572.0	0.0	57.0	45.0
JAWA TIMUR	517574	200209	52445	971027	292169	83538	2611658	106417	4079350	156703.0	...	405513	551285	58148	98601	73982	188639	368149.0	1548312.0	1843.0	2482.0
BANTEN	5866	4548	9878	90105	17618	19643	2411	515	94778	22488.0	...	92262	3830	9077	11994	36096	975	4788.0	0.0	18.0	23.0
BALI	11263	1596	5465	74130	10924	6065	861555	5526	161993	59340.0	...	58417	128845	10313	1472	1186	803	53426.0	6.0	33207.0	1155.0
NUSA TENGGARA BARAT	33294	1996	2471	102770	44368	7947	19847	3325	435121	62272.0	...	55385	262	41150	84638	1761	7502	63944.0	772.0	1312.0	496.0
NUSA TENGGARA TIMUR	70890	2149	27	9897	32179	5539	160158	6106	229802	73.0	...	18758	8604	3253	11768	10532	1035	6703.0	108.0	42.0	352.0
KALIMANTAN BARAT	4548	4075	20404	74767	13412	8950	324928	1978	23131	17111.0	...	37101	5498	13112	5430	5948	1169	16216.0	0.0	30.0	0.0
KALIMANTAN TENGAH	1844	3274	4014	18952	10389	6626	21798	1088	6871	1280.0	...	31475	2880	4450	2590	4011	2470	40153.0	0.0	25.0	312.0
KALIMANTAN SELATAN	120	2388	20608	54851	5685	2545	387016	1293	22399	3589.0	...	45627	2664	4846	4094	6805	2720	47229.0	0.0	1.0	0.0
KALIMANTAN TIMUR	2460	3342	17592	32710	5834	4380	31573	1407	12403	938.0	...	31241	34227	6819	3068	4925	1457	14669.0	0.0	0.0	0.0
KALIMANTAN UTARA	261	661	13196	23194	1119	1986	15872	285	20784	281.0	...	24914	13257	332	1204	2053	1687	4647.0	0.0	2.0	0.0
SULAWESI UTARA	17925	1304	21625	38768	8585	3403	3266	1299	48286	6371.0	...	28672	37950	2	5702	3070	2413	2420.0	0.0	2.0	0.0
SULAWESI TENGAH	17891	763	15185	116981	2195	3044	10056	666	37718	5294.0	...	28801	2535	574	1100	1163	6594	19115.0	0.0	128.0	3.0
SULAWESI SELATAN	24000	1523	67959	167739	36423	5804	30905	106276	335685	9524.0	...	108485	34876	2064	5400	40360	2309	25453.0	0.0	10.0	279.0
SULAWESI TENGGARA	3275	1731	11905	48369	5650	4700	92361	3570	47350	3667.0	...	37403	5473	757	3662	8024	772	3626.0	0.0	119.0	24.0
GORONTALO	121	47	1026	14641	66	16	62034	335	8747	9.0	...	4100	30	1	102	19	142	394.0	0.0	0.0	0.0
SULAWESI BARAT	1148	294	23170	54973	936	807	54884	592	43288	3835.0	...	18296	224	182	533	3308	72	589.0	0.0	0.0	0.0
MALUKU	4442	1551	8460	36085	3768	3596	47175	2595	18860	3192.0	...	3549	6886	60	1933	9094	966	3067.0	4.0	0.0	0.0
MALUKU UTARA	2674	725	9582	15432	1684	1959	7184	692	14102	2971.0	...	20983	6297	24	1135	2294	329	2840.0	0.0	1.0	0.0
PAPUA BARAT	10375	1785	14385	10675	27576	1865	13989	1661	16284	42.0	...	13179	7575	1285	2645	4132	796	2316.0	0.0	0.0	1.0
PAPUA	1233	559	432	1749	1622	855	80932	280	14168	267.0	...	2261	324	114	626	956	6667	16165.0	0.0	2.0	0.0

	Alpukat	Belimbing	Duku	Durian	Jambu Biji	Jambu Air	Jeruk Siam	Jeruk Besar	Mangga	Manggis	...	Rambutan	Salak	Sawo	Sirsak	Sukun	Melon	Semangka	Apel	Anggur	Stroberi
Provinsi
ACEH	0.291363	0.021293	0.161580	0.133912	0.074703	0.096644	0.017521	0.290875	0.017983	0.038519	...	0.226465	0.002106	0.173224	0.034549	0.043912	0.005694	0.057357	0.000194	0.000000	0.003780
SUMATERA UTARA	0.227616	0.096422	0.424314	0.322584	0.131779	0.312034	0.452902	0.048885	0.029879	0.322565	...	0.178499	0.602756	0.795422	0.090346	0.054170	0.027122	0.306789	0.000000	0.000181	0.006715
SUMATERA BARAT	0.514969	0.014503	0.148561	0.537034	0.090682	0.188195	0.145713	0.015408	0.010812	1.000000	...	0.196029	0.004290	0.466358	0.273952	0.083238	0.029693	0.138770	0.000030	0.000120	0.011579
RIAU	0.008617	0.033268	0.157787	0.076951	0.107685	0.185115	0.097217	0.044514	0.011785	0.140105	...	0.200502	0.008661	0.200015	0.045899	0.081517	0.019472	0.174503	0.000000	0.001777	0.000000
JAMBI	0.063911	0.017646	0.365544	0.065753	0.037509	0.071546	0.064152	0.007737	0.003035	0.138517	...	0.041989	0.012143	0.150015	0.060194	0.076810	0.014309	0.045897	0.000000	0.000361	0.000555
SUMATERA SELATAN	0.203494	0.020089	1.000000	0.120168	0.060922	0.194117	0.065724	0.010708	0.011441	0.026759	...	0.088811	0.004215	0.272180	0.042610	0.073144	0.027891	0.123369	0.000002	0.000361	0.005129
BENGKULU	0.042578	0.005206	0.017671	0.056904	0.016235	0.021493	0.026249	0.012456	0.004073	0.072117	...	0.038271	0.007032	0.071056	0.013310	0.014827	0.007199	0.007657	0.000031	0.001114	0.004203
LAMPUNG	0.122933	0.050719	0.214332	0.063845	0.154097	0.202906	0.068819	0.032377	0.013787	0.126812	...	0.130913	0.018590	0.567609	0.087107	0.178575	0.011186	0.170499	0.000000	0.002710	0.000026
KEP. BANGKA BELITUNG	0.007672	0.001729	0.014998	0.012514	0.005442	0.031046	0.001847	0.002200	0.000000	0.010851	...	0.008516	0.002080	0.022865	0.003432	0.013243	0.005370	0.027209	0.000000	0.000000	0.000000
KEP. RIAU	0.000696	0.001449	0.006256	0.011251	0.004635	0.006608	0.000019	0.000226	0.000050	0.002551	...	0.006444	0.014118	0.034617	0.010071	0.028965	0.001415	0.014027	0.000000	0.000000	0.000000
DKI JAKARTA	0.006383	0.038114	0.000051	0.000000	0.014546	0.041064	0.000000	0.000000	0.003484	0.000000	...	0.020258	0.000079	0.008731	0.001259	0.021401	0.000000	0.000000	0.000000	0.000422	0.000000
JAWA BARAT	0.649173	0.125903	0.165979	0.246797	0.740837	0.507414	0.058278	0.218751	0.326976	0.717383	...	1.000000	0.056649	1.000000	0.430664	0.649012	0.013227	0.080617	0.000000	0.005059	1.000000
JAWA TENGAH	0.540239	0.289071	0.330441	0.492965	1.000000	1.000000	0.032106	0.522158	0.353382	0.130046	...	0.819284	1.000000	0.808460	0.648321	1.000000	0.434302	0.553428	0.000043	0.011714	0.076614
DI YOGYAKARTA	0.047573	0.011126	0.062873	0.024968	0.051108	0.076174	0.002677	0.004597	0.020729	0.036384	...	0.089875	0.116884	0.110593	0.086031	0.331351	0.232048	0.055880	0.000000	0.001717	0.001190
JAWA TIMUR	1.000000	1.000000	0.378685	1.000000	0.946727	0.540543	1.000000	1.000000	1.000000	0.708794	...	0.928405	0.386380	0.741113	1.000000	0.636465	1.000000	1.000000	1.000000	0.055500	0.065616
BANTEN	0.011104	0.022487	0.071167	0.092186	0.056887	0.127023	0.000675	0.004418	0.021815	0.101717	...	0.207209	0.002663	0.115678	0.120732	0.310449	0.005116	0.013006	0.000000	0.000542	0.000608
BALI	0.021534	0.007739	0.039286	0.075723	0.035192	0.039148	0.329722	0.051527	0.038315	0.268405	...	0.129288	0.090288	0.131432	0.013909	0.010042	0.004204	0.145121	0.000004	1.000000	0.030535
NUSA TENGGARA BARAT	0.064110	0.009737	0.017656	0.105237	0.143586	0.051328	0.007353	0.030835	0.105366	0.281667	...	0.122307	0.000163	0.524465	0.858242	0.014990	0.039718	0.173691	0.000499	0.039510	0.013113
NUSA TENGGARA TIMUR	0.136766	0.010501	0.000000	0.009529	0.104081	0.035744	0.061091	0.056979	0.054962	0.000330	...	0.037981	0.006010	0.041448	0.118438	0.090466	0.005434	0.018207	0.000070	0.001265	0.009306
KALIMANTAN BARAT	0.008557	0.020124	0.147210	0.076380	0.043255	0.057820	0.124197	0.018172	0.004226	0.077396	...	0.080212	0.003833	0.167106	0.054092	0.051020	0.006144	0.044047	0.000000	0.000903	0.000000
KALIMANTAN TENGAH	0.003332	0.016122	0.028803	0.018861	0.033458	0.042779	0.008100	0.009805	0.000234	0.005790	...	0.067259	0.001998	0.056705	0.025259	0.034352	0.013041	0.109067	0.000000	0.000753	0.008248
KALIMANTAN SELATAN	0.000000	0.011696	0.148684	0.055856	0.018212	0.016367	0.147976	0.011732	0.004046	0.016234	...	0.099841	0.001846	0.061752	0.040528	0.058395	0.014367	0.128288	0.000000	0.000030	0.000000
KALIMANTAN TIMUR	0.004522	0.016462	0.126895	0.033039	0.018694	0.028243	0.011844	0.012804	0.001592	0.004243	...	0.066721	0.023969	0.086899	0.030112	0.042217	0.007671	0.039845	0.000000	0.000000	0.000000
KALIMANTAN UTARA	0.000272	0.003068	0.095137	0.023232	0.003413	0.012750	0.005830	0.002256	0.003650	0.001271	...	0.052154	0.009271	0.004219	0.011188	0.017503	0.008890	0.012623	0.000000	0.000060	0.000000
SULAWESI UTARA	0.034409	0.006280	0.156031	0.039282	0.027611	0.021920	0.001002	0.011789	0.010401	0.028817	...	0.060806	0.026578	0.000013	0.056853	0.026254	0.012739	0.006573	0.000000	0.000060	0.000000
SULAWESI TENGAH	0.034343	0.003577	0.109507	0.119882	0.006900	0.019597	0.003603	0.005838	0.007807	0.023946	...	0.061103	0.001756	0.007303	0.010132	0.009844	0.034904	0.051922	0.000000	0.003855	0.000079
SULAWESI SELATAN	0.046149	0.007374	0.490764	0.172190	0.117836	0.037459	0.011588	0.998674	0.080956	0.043079	...	0.244559	0.024424	0.026294	0.053787	0.347142	0.012188	0.069138	0.000000	0.000301	0.007376
SULAWESI TENGGARA	0.006097	0.008413	0.085811	0.049176	0.018098	0.030314	0.035125	0.033138	0.010171	0.016586	...	0.080907	0.003815	0.009636	0.036142	0.068885	0.004040	0.009849	0.000000	0.003584	0.000634
GORONTALO	0.000002	0.000000	0.007217	0.014418	0.000000	0.000000	0.023510	0.002726	0.000695	0.000041	...	0.004234	0.000000	0.000000	0.000000	0.000000	0.000700	0.001070	0.000000	0.000000	0.000000
SULAWESI BARAT	0.001987	0.001234	0.167193	0.055981	0.002820	0.005119	0.020772	0.005142	0.009174	0.017346	...	0.036917	0.000136	0.002307	0.004376	0.028302	0.000329	0.001600	0.000000	0.000000	0.000000
MALUKU	0.008352	0.007514	0.060923	0.036517	0.011998	0.023169	0.017819	0.023972	0.003177	0.014438	...	0.002965	0.004805	0.000752	0.018589	0.078092	0.005068	0.008331	0.000003	0.000000	0.000000
MALUKU UTARA	0.004936	0.003387	0.069029	0.015233	0.005244	0.012575	0.002503	0.006082	0.002009	0.013438	...	0.043104	0.004393	0.000293	0.010487	0.019577	0.001691	0.007714	0.000000	0.000030	0.000000
PAPUA BARAT	0.019818	0.008683	0.103727	0.010331	0.089162	0.011966	0.005109	0.015192	0.002545	0.000190	...	0.025136	0.005288	0.016365	0.025818	0.035393	0.004167	0.006291	0.000000	0.000000	0.000026
PAPUA	0.002151	0.002558	0.002926	0.001133	0.005043	0.005430	0.030748	0.002209	0.002025	0.001208	...	0.000000	0.000206	0.001440	0.005320	0.008063	0.035291	0.043909	0.000000	0.000060	0.000000

	principal component 1	principal component 2	principal component 3
Provinsi
ACEH	-0.092806	0.047337	0.024086
SUMATERA UTARA	0.472903	-0.261143	-0.021415
SUMATERA BARAT	0.384959	-0.152729	-0.776127
RIAU	-0.083750	-0.117560	0.058292
JAMBI	-0.239929	-0.039618	0.026891
SUMATERA SELATAN	0.007483	-0.239554	0.309540
BENGKULU	-0.411179	0.058462	-0.062127
LAMPUNG	0.349867	-0.437603	0.146488
KEP. BANGKA BELITUNG	-0.482492	0.072243	-0.000884
KEP. RIAU	-0.487729	0.066953	0.001912
DKI JAKARTA	-0.489489	0.075747	0.004065
JAWA BARAT	1.596476	-1.023694	-0.552769
JAWA TENGAH	2.174177	-0.537051	0.835334
DI YOGYAKARTA	-0.193676	0.018822	0.118851
JAWA TIMUR	3.453580	1.080844	-0.173608
BANTEN	-0.182623	-0.058624	0.004994
BALI	-0.156027	0.187157	-0.351779
NUSA TENGGARA BARAT	0.220100	-0.113325	-0.286556
NUSA TENGGARA TIMUR	-0.192434	0.107810	-0.000499
KALIMANTAN BARAT	-0.261015	-0.015714	-0.004639
KALIMANTAN TENGAH	-0.390270	0.060895	0.019443
KALIMANTAN SELATAN	-0.319736	0.101512	0.022756
KALIMANTAN TIMUR	-0.355860	0.032815	0.048697
KALIMANTAN UTARA	-0.427224	0.083359	0.032041
SULAWESI UTARA	-0.397065	0.071251	0.027039
SULAWESI TENGAH	-0.409109	0.090917	0.005913
SULAWESI SELATAN	0.106292	0.246762	0.437292
SULAWESI TENGGARA	-0.393085	0.076199	0.025919
GORONTALO	-0.513739	0.096608	-0.010166
SULAWESI BARAT	-0.450433	0.080170	0.024656
MALUKU	-0.443985	0.086514	0.024672
MALUKU UTARA	-0.479259	0.074951	0.009942
PAPUA BARAT	-0.420570	0.067151	0.035231
PAPUA	-0.492354	0.112136	-0.003484

1.Import Data¶

2.Explanatory Data Analysis (EDA)¶

2.1 Data Cleaning¶

2.2 Data Visualization¶

2.3 Data Transformation¶

2.4 VIF¶

3. PCA¶

4. K-Means¶

k=3¶

k=4¶

k=5¶

5. Single Linkage¶

Combination K-Means & Single Linkage¶

6. Evaluasi cluster¶

6.1 Calinski Harabasz Index¶

6.2 Silhouette coefficient¶

Original Data¶

6. Evaluasi cluster¶

6.1 Calinski Harabasz Index¶

6.2 Silhouette coefficient¶

6.3 Visualisasi¶

6.3.1 K-Means++¶

6.3.2 Random¶

6.3.3 Hierarchical KMeans Clustering¶

6.3.4 Single Linkage¶

	Variables	VIF
0	Alpukat	231.974512
1	Belimbing	1400.773457
2	Duku	13.171865
3	Durian	257.046865
4	Jambu Biji	359.296810
5	Jambu Air	1813.794145
6	Jeruk Siam	150.844569
7	Jeruk Besar	111.155240
8	Mangga	4794.463991
9	Manggis	106.549009
10	Nangka	186.399033
11	Nanas	155.402171
12	Pepaya	56.890755
13	Pisang	411.241944
14	Rambutan	309.253186
15	Salak	219.163682
16	Sawo	566.568867
17	Sirsak	442.691592
18	Sukun	342.950252
19	Melon	334.925102
20	Semangka	205.517478
21	Apel	2721.030432
22	Anggur	13.983284
23	Stroberi	428.268248

	PC1	PC2	PC3
Alpukat (Ton)	0.252451	-0.022119	-0.181361
Belimbing (Ton)	0.183705	0.242558	0.001300
Duku (Ton)	0.086611	-0.070551	0.270380
Durian (Ton)	0.216049	0.135159	-0.109172
Jambu Biji (Ton)	0.284728	-0.118264	0.114043
Jambu Air (Ton)	0.211105	-0.228374	0.190181
Jeruk Siam (Ton)	0.157324	0.292070	-0.170702
Jeruk Besar (Ton)	0.213581	0.266015	0.256476
Mangga (Ton)	0.203846	0.181947	-0.030309
Manggis (Ton)	0.184538	-0.102790	-0.599328
Nangka (Ton)	0.256665	0.015693	0.125663
Nenas (Ton)	0.125161	-0.270899	0.192270
Pepaya (Ton)	0.221634	0.077304	-0.010064
Pisang (Ton)	0.222451	0.029697	-0.068860
Rambutan (Ton)	0.276040	-0.167025	-0.032217
Salak (Ton)	0.163654	-0.101400	0.330312
Sawo (Ton)	0.277351	-0.412929	-0.167949
Sirsak (Ton)	0.245180	0.038517	-0.120927
Sukun (Ton)	0.222552	-0.169971	0.253015
Melon (Ton)	0.186388	0.266910	0.087919
Semangka (Ton)	0.207232	0.174681	0.068631
Apel (Ton)	0.148446	0.349422	-0.079629
Anggur (Ton)	0.003172	0.074561	-0.167209
Stroberi (Ton)	0.085463	-0.320293	-0.237788